This chapter is concerned with estimating the distribution of a random variable U when one observes realizations not of U but of W = U +ε, where ε is random variable that is independent of U. Such estimation problems are called deconvolution problems because the distribution of the observed random variable, W, is the convolution of the distributions of U and ε. Estimating the distribution of U requires deconvoluting the distribution of the observed random variable W.
KeywordsCharacteristic Function Consistent Estimator Nonparametric Estimator Monte Carlo Experiment Deconvolution Problem
Unable to display preview. Download preview PDF.